The Agent Skills Directory

[DYNAMIC_CONTEXT_INJECTION]: The skill utilizes shell-based context injection via !find commands to automatically list agent definition files and settings files. These commands are used for environment discovery within the project's agent configuration directory and do not involve network access or sensitive file exposure.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze third-party agent configuration files. While this presents an indirect prompt injection surface where instructions in audited files could influence the auditing agent, the risk is minimal given the read-only audit workflow.
Ingestion points: Agent definition files and settings.local.json identified via discovery commands in SKILL.md.
Boundary markers: None explicitly instructed for the processing of configuration files.
Capability inventory: File discovery and reading (Glob, Read) and status tracking (TodoWrite) as defined in allowed-tools.
Sanitization: None identified for the analysis of file content.

meta-audit